Matching (graph theory)

In the mathematical discipline of graph theory, a matching or independent edge set in a graph is a set of edges
without common vertices. It may also be an entire graph
consisting of edges without common vertices. Bipartite
matching is a special case of a network ow problem.

some literature, the term complete matching is used. In

the above gure, only part (b) shows a perfect matching.
A perfect matching is also a minimum-size edge cover.
Thus, (G) (G) , that is, the size of a maximum matching is no larger than the size of a minimum edge cover.

A near-perfect matching is one in which exactly one

vertex is unmatched. This can only occur when the graph
1 Denition
has an odd number of vertices, and such a matching must
be maximum. In the above gure, part (c) shows a nearGiven a graph G = (V,E), a matching M in G is a set of perfect matching. If, for every vertex in a graph, there is
pairwise non-adjacent edges; that is, no two edges share a near-perfect matching that omits only that vertex, the
graph is also called factor-critical.
a common vertex.
A vertex is matched (or saturated) if it is an endpoint of Given a matching M,
one of the edges in the matching. Otherwise the vertex is
an alternating path is a path in which the edges
unmatched.
belong alternatively to the matching and not to the
A maximal matching is a matching M of a graph G with
matching.
the property that if any edge not in M is added to M, it is
no longer a matching, that is, M is maximal if it is not a
an augmenting path is an alternating path that
proper subset of any other matching in graph G. In other
starts from and ends on free (unmatched) vertices.
words, a matching M of a graph G is maximal if every
edge in G has a non-empty intersection with at least one One can prove that a matching is maximum if and only
edge in M. The following gure shows examples of max- if it does not have any augmenting path. (This result is
imal matchings (red) in three graphs.
sometimes called Berges lemma.)

2 Properties
In any graph without isolated vertices, the sum of the
matching number and the edge covering number equals
A maximum matching (also known as maximumthe number of vertices.[2] If there is a perfect matching,
cardinality matching[1] ) is a matching that contains the
then both the matching number and the edge cover numlargest possible number of edges. There may be many
ber are |V| / 2.
maximum matchings. The matching number (G) of a
graph G is the size of a maximum matching. Note that ev- If A and B are two maximal matchings, then |A| 2|B|
ery maximum matching is maximal, but not every maxi- and |B| 2|A|. To see this, observe that each edge in B \
mal matching is a maximum matching. The following g- A can be adjacent to at most two edges in A \ B because
ure shows examples of maximum matchings in the same A is a matching; moreover each edge in A \ B is adjacent
to an edge in B \ A by maximality of B, hence
three graphs.
(a)

(b)

(c)

|A \ B| 2|B \ A|.
Further we get that

A perfect matching (a.k.a. 1-factor) is a matching which

matches all vertices of the graph. That is, every vertex of
the graph is incident to exactly one edge of the matching.
Figure (b) above is an example of a perfect matching. Every perfect matching is maximum and hence maximal. In

|A| = |A B| + |A \ B| 2|B A| + 2|B \ A| = 2|B|.

In particular, this shows that any maximal matching is a
2-approximation of a maximum matching and also a 2approximation of a minimum maximal matching. This
1

ALGORITHMS AND COMPUTATIONAL COMPLEXITY

inequality is tight: for example, if G is a path with 3 edges 4.2 In weighted bipartite graphs
and 4 nodes, the size of a minimum maximal matching is
1 and the size of a maximum matching is 2.
In a weighted bipartite graph, each edge has an associated
value. A maximum weighted bipartite matching[3] is
dened as a matching where the sum of the values of the
edges in the matching have a maximal value. If the graph
3 Matching polynomials
is not complete bipartite, missing edges are inserted with
value zero. Finding such a matching is known as the
Main article: Matching polynomial
assignment problem. The remarkable Hungarian algorithm solves the assignment problem and it was one of
A generating function of the number of k-edge matchings the beginnings of combinatorial optimization algorithms.
in a graph is called a matching polynomial. Let G be a It uses a modied shortest path search in the augmenting
graph and mk be the number of k-edge matchings. One path algorithm. If the BellmanFord algorithm is used
for this step, the running time of the Hungarian algorithm
matching polynomial of G is
becomes O(V 2 E) , or the edge cost can be shifted with
a potential to achieve O(V 2 log V + V E) running time
with the Dijkstra algorithm and Fibonacci heap.[7]

mk x k .
k0

Another denition gives the matching polynomial as

4.3 In general graphs

Main article: Edmondss matching algorithm

(1)k mk xn2k ,

There is a polynomial time algorithm to nd a maximum

matching or a maximum weight matching in a graph that
where n is the number of vertices in the graph. Each is not bipartite; it is due to Jack Edmonds, is called the
type has its uses; for more information see the article on paths, trees, and owers method or simply Edmondss algorithm, and uses bidirected edges. A generalization of
matching polynomials.
the same technique can also be used to nd maximum
independent sets in claw-free graphs. Edmonds algorithm has subsequently been improved to run in time
4 Algorithms and computational O(VE) time, matching the time for bipartite maximum
matching.[8]
k0

complexity

4.1

In unweighted bipartite graphs

Another (randomized) algorithm by Mucha and

Sankowski,[4] based on the fast matrix multiplication
algorithm, gives O(V 2.376 ) complexity.

Matching problems are often concerned with bipartite

graphs. Finding a maximum bipartite matching[3] (of- 4.4 Maximal matchings
ten called a maximum cardinality bipartite matching)
in a bipartite graph G = (V = (X, Y ), E) is perhaps the
A maximal matching can be found with a simple greedy
simplest problem.
algorithm. A maximum matching is also a maximal
The augmenting path algorithm nds it by nding an aug- matching, and hence it is possible to nd a largest
menting path from each x X to Y and adding it to the maximal matching in polynomial time. However, no
matching if it exists. As each path can be found in O(E) polynomial-time algorithm is known for nding a mintime, the running time is O(V E) . This solution is equiv- imum maximal matching, that is, a maximal matching
alent to adding a super source s with edges to all vertices that contains the smallest possible number of edges.
in X , and a super sink t with edges from all vertices in Y Note that a maximal matching with k edges is an edge
, and nding a maximal ow from s to t . All edges with dominating set with k edges. Conversely, if we are given
ow from X to Y then constitute a maximum matching. a minimum edge dominating set with k edges, we can
An improvement over this
is the HopcroftKarp algorithm, which runs in O( V E) time. Another approach
is based on the fast matrix multiplication algorithm and
gives O(V 2.376 ) complexity,[4] which is better in theory
for suciently dense graphs, but in practice the algorithm
10/7 ) is posis slower.[5] Finally, for sparse graphs, O(E
sible with Madrys algorithm based on electric ows. [6]

construct a maximal matching with k edges in polynomial time. Therefore the problem of nding a minimum
maximal matching is essentially equal to the problem of
nding a minimum edge dominating set.[9] Both of these
two optimisation problems are known to be NP-hard; the
decision versions of these problems are classical examples of NP-complete problems.[10] Both problems can be

3
approximated within factor 2 in polynomial time: simply Halls marriage theorem provides a characterization of
nd an arbitrary maximal matching M.[11]
bipartite graphs which have a perfect matching and the
Tutte theorem provides a characterization for arbitrary
graphs.

4.5

Counting problems

Main article: Hosoya index

A perfect matching is a spanning 1-regular subgraph,

a.k.a. a 1-factor. In general, a spanning k-regular subgraph is a k-factor.

The number of matchings in a graph is known as the

Hosoya index of the graph. It is #P-complete to compute
6 Applications
this quantity. It remains #P-complete in the special case
of counting the number of perfect matchings in a given
bipartite graph, because computing the permanent of an 6.1 Matching in general graphs
arbitrary 01 matrix (another #P-complete problem) is
A Kekul structure of an aromatic compound conthe same as computing the number of perfect matchsists of a perfect matching of its carbon skeleings in the bipartite graph having the given matrix as its
ton, showing the locations of double bonds in the
biadjacency matrix. However, there exists a fully polynochemical structure. These structures are named afmial time randomized approximation scheme for countter Friedrich August Kekul von Stradonitz, who
ing the number of bipartite matchings.[12] A remarkable
showed that benzene (in graph theoretical terms, a
theorem of Kasteleyn states that the number of perfect
6-vertex cycle) can be given such a structure.[19]
matchings in a planar graph can be computed exactly in
polynomial time via the FKT algorithm.
The Hosoya index is the number of non-empty
The number of perfect matchings in a complete graph Kn
matchings plus one; it is used in computational
(with n even) is given by the double factorial (n 1)!!.[13]
chemistry and mathematical chemistry investigaThe numbers of matchings in complete graphs, without
tions for organic compounds.
constraining the matchings to be perfect, are given by the
telephone numbers.[14]

6.2 Matching in Bipartite Graphs

4.6

maximally-matchable

Graduation problem is about choosing minimum set

of classes from given requirements for graduation.

One of the basic problems in matching theory is to nd

in a given graph all edges that may be extended to a maximum matching in the graph. (Such edges are called
maximally-matchable edges, or allowed edges.) The
best deterministic algorithm for solving this problem in
general graphs runs in time O(V E) .[15] There exists a
randomized algorithm that solves this problem in time
2.376 ) .[16] In the case of bipartite graphs, it is posO(V
sible to nd a single maximum matching and then use
it in order to nd all maximally-matchable edges in linear time;[17] the resulting overall runtime is O(V 1/2 E)
for general bipartite graphs and O((V / log V )1/2 E) for
dense bipartite graphs with E = (V 2 ) . In cases where
one of the maximum matchings is known upfront,[18] the
overall runtime of the algorithm is O(V + E) .

Hitchcock Transport Problem involves bipartite

matching as sub-problem.

Finding
edges

all

Characterizations and notes

Knigs theorem states that, in bipartite graphs, the

maximum matching is equal in size to the minimum
vertex cover. Via this result, the minimum vertex cover,
maximum independent set, and maximum vertex biclique
problems may be solved in polynomial time for bipartite
graphs.

Subtree isomorphism problem involves bipartite

matching as sub-problem.

7 See also
DulmageMendelsohn decomposition, a partition of
the vertices of a bipartite graph into subsets such that
each edge belongs to a perfect matching if and only
if its endpoints belong to the same subset
Edge coloring, a partition of the edges of a graph
into matchings
Matching preclusion, the minimum number of edges
to delete to prevent a perfect matching from existing
Rainbow matching, a matching in an edge-colored
bipartite graph with no repeated colors
Skew-symmetric graph, a type of graph that can be
used to model alternating path searches for matchings

9 FURTHER READING
Stable matching, a matching in which no two ele- [12] Bezkov, Ivona; tefankovi, Daniel; Vazirani, Vijay V.;
Vigoda, Eric (2008). Accelerating Simulated Annealments prefer each other to their matched partners
Vertex independent set, a set of vertices (rather than
edges) no two of which are adjacent to each other

References

[1] Alan Gibbons, Algorithmic Graph Theory, Cambridge

University Press, 1985, Chapter 5.
[2] Gallai, Tibor (1959), "ber extreme Punkt- und Kantenmengen, Ann. Univ. Sci. Budapest. Etvs Sect. Math.
2: 133138.
[3] West, Douglas Brent (1999), Introduction to Graph Theory (2nd ed.), Prentice Hall, Chapter 3, ISBN 0-13014400-2
[4] Mucha, M.; Sankowski, P. (2004), Maximum Matchings
via Gaussian Elimination (PDF), Proc. 45th IEEE Symp.
Foundations of Computer Science, pp. 248255
[5] Chandran, Bala G.; Hochbaum, Dorit S. (2011), Practical and theoretical improvements for bipartite matching using the pseudoow algorithm, arXiv:1105.1569, the theoretically ecient algorithms listed above tend to perform
poorly in practice.
[6] Madry, A (2013), Navigating Central Path with Electrical Flows: From Flows to Matchings, and Back, Foundations of Computer Science (FOCS), 2013 IEEE 54th Annual
Symposium on, pp. 253262, arXiv:1307.2205
[7] Fredman, Michael L.; Tarjan, Robert Endre (1987), Fibonacci heaps and their uses in improved network optimization algorithms, Journal of the ACM 34 (3): 596
615, doi:10.1145/28869.28874

[8] Micali, S.; Vazirani, V. V. (1980), An O( |V ||E|) algorithm for nding maximum matching in general graphs,
Proc. 21st IEEE Symp. Foundations of Computer Science,
pp. 1727, doi:10.1109/SFCS.1980.12.
[9] Yannakakis, Mihalis; Gavril, Fanica (1980), Edge dominating sets in graphs, SIAM Journal on Applied Mathematics 38 (3): 364372, doi:10.1137/0138030.
[10] Garey, Michael R.; Johnson, David S. (1979), Computers
and Intractability: A Guide to the Theory of NPCompleteness, W.H. Freeman, ISBN 0-7167-1045-5.
Edge dominating set (decision version) is discussed under
the dominating set problem, which is the problem GT2 in
Appendix A1.1. Minimum maximal matching (decision
version) is the problem GT10 in Appendix A1.1.
[11] Ausiello, Giorgio; Crescenzi, Pierluigi; Gambosi, Giorgio; Kann, Viggo; Marchetti-Spaccamela, Alberto; Protasi, Marco (2003), Complexity and Approximation: Combinatorial Optimization Problems and Their Approximability Properties, Springer. Minimum edge dominating set
(optimisation version) is the problem GT3 in Appendix B
(page 370). Minimum maximal matching (optimisation
version) is the problem GT10 in Appendix B (page 374).
See also Minimum Edge Dominating Set and Minimum
Maximal Matching in the web compendium.

ing for the Permanent and Combinatorial Counting Problems. SIAM Journal on Computing 37 (5): 14291454.
doi:10.1137/050644033.
[13] Callan, David (2009), A combinatorial survey of identities
for the double factorial, arXiv:0906.1317.
[14] Tichy, Robert F.; Wagner, Stephan (2005), Extremal
problems for topological indices in combinatorial chemistry (PDF), Journal of Computational Biology 12 (7):
10041013, doi:10.1089/cmb.2005.12.1004.
[15] de Carvalho, Marcelo H.; Cheriyan, Joseph (2005), An
O(V E) algorithm for ear decompositions of matchingcovered graphs, Proc. ACM/SIAM Symposium on Discrete
Algorithms (SODA), pp. 415423.
[16] Rabin, Michael O.; Vazirani, Vijay V. (1989), Maximum matchings in general graphs through randomization, J. of Algorithms 10: 557567, doi:10.1016/01966774(89)90005-9.
[17] Tassa, Tamir (2012), Finding all maximally-matchable
edges in a bipartite graph, Theoretical Computer Science
423: 5058, doi:10.1016/j.tcs.2011.12.071.
[18] Gionis, Aris; Mazza, Arnon; Tassa, Tamir (2008), "kAnonymization revisited, International Conference on
Data Engineering (ICDE), pp. 744753.
[19] See, e.g., Trinajsti, Nenad; Klein, Douglas J.; Randi,
Milan (1986), On some solved and unsolved problems of chemical graph theory", International Journal of Quantum Chemistry 30 (S20): 699742,
doi:10.1002/qua.560300762.

9 Further reading
1. Lszl Lovsz; M. D. Plummer (1986), Matching
Theory, North-Holland, ISBN 0-444-87916-1
2. Thomas H. Cormen, Charles E. Leiserson, Ronald
L. Rivest and Cliord Stein (2001), Introduction to
Algorithms (second ed.), MIT Press and McGraw
Hill, Chapter 26, pp. 643700, ISBN 0-262-531968
3. Andrs Frank (2004). On Kuhns Hungarian Method
A tribute from Hungary (PDF) (Technical report).
Egervry Research Group.
4. Michael L. Fredman and Robert E. Tarjan (1987),
Fibonacci heaps and their uses in improved network optimization algorithms, Journal of the ACM
34 (3): 595615, doi:10.1145/28869.28874.
5. S. J. Cyvin and Ivan Gutman (1988), Kekule Structures in Benzenoid Hydrocarbons, Springer-Verlag
6. Marek Karpinski and Wojciech Rytter (1998), Fast
Parallel Algorithms for Graph Matching Problems,
Oxford University Press, ISBN 978-0-19-850162-6

10

External links

A graph library with HopcroftKarp and Push

Relabel-based maximum cardinality matching implementation

11

11
11.1

TEXT AND IMAGE SOURCES, CONTRIBUTORS, AND LICENSES

Text and image sources, contributors, and licenses

Text

Matching (graph theory) Source: https://en.wikipedia.org/wiki/Matching_(graph_theory)?oldid=666001845 Contributors: Edemaine,

Jdpipe, Tomo, Michael Hardy, Mac, Ronz, TheEternalVortex, Dcoetzee, Gutza, Omegatron, Altenmann, MathMartin, Giftlite,
Mintleaf~enwiki, Kainaw, DonDiego, Zaslav, Aranel, El C, Jammydodger, Squizzz~enwiki, LOL, Oliphaunt, Adking80, Mamling,
Rjwilmsi, MarSch, Eubot, Mathbot, Maxal, Chobot, Michael Slone, Kerry Raymond, Nils Grimsmo, Arthur Rubin, H@r@ld, RobertBorgersen, SmackBot, Rentier, Bluebot, Mgccl, Daqu, Oudmatie, Mental Blank, Lambiam, MickPurcell, RamanTheGreat, Ylloh, Mmarci1111,
CRGreathouse, Cyhawk, Sytelus, Green Snake, Thijs!bot, Itamarro, Hermel, David Eppstein, Azer Red, Nono64, Cmansley, Madhero88,
Aednichols, Debamf, Kamyar1, ClueBot, Excirial, Tim32, Bender2k14, SchreiberBike, Fghtngthfght, Addbot, DOI bot, Fyrael, Smartnut007, Favonian, Movado73, Luckas-bot, Yobot, Kilom691, Calle, AnomieBOT, Erel Segal, Mangarah, Citation bot, Xqbot, Gilo1969,
Miym, LucienBOT, Pparys, , RedBot, Markoid, RjwilmsiBot, Alph Bot, EmausBot, Kraymer, WikitanvirBot, ZroBot, Mastergreg82, Surngpete~enwiki, Wcherowi, Faisalsyn, Helpful Pixie Bot, Sahba ezami, Ankushshah89, Shurakai, Scottkwong, Rahul38888,
Pmendek, KasparBot, Ahaider3 and Anonymous: 74

11.2

Images

File:Maximal-matching.svg Source: https://upload.wikimedia.org/wikipedia/commons/e/e1/Maximal-matching.svg License: CC BYSA 3.0 Contributors: Own work Original artist: Miym
File:Maximum-matching-labels.svg Source: https://upload.wikimedia.org/wikipedia/commons/9/98/Maximum-matching-labels.svg
License: CC BY-SA 3.0 Contributors: Own work Original artist: Miym

11.3

Content license

Creative Commons Attribution-Share Alike 3.0